Using mlpack in Databricks how-to

getstarted.md

layout	title	displaytitle	description	order
default-nav	Get Started	Get Started	mlpack | fast, flexible machine learning library in C++	2

Get Started

Overview

There are multiple ways to get mlpack up and running. Python bindings can be installed using pip or conda, or built from source (see "Build from source" tutorials). Julia bindings can be installed via Julia's Pkg package manager.

For C++, if mlpack is not available via your preferred OS package manager, or if you need to build your own version (e.g. to apply optimizations, use a different set of BLAS/LAPACK, or build a different configuration), please refer to the "Build from source" tutorials. For Windows, prebuilt binaries will help you get started without the need of building mlpack. These packages include both the C++ mlpack library as well as the CLI tools.

Once you get mlpack running, check out the documentation or the examples repository, which contains simple example usages of mlpack.

Here is a summary of the currently available distribution options you can use depending on your needs:

Python

Get via pip or conda

pip install mlpack

conda install -c conda-forge mlpack

• mlpack in Python quickstart guide

Databricks

• From your workspace dropdown, select Create -> Library. Then specify PyPI and use "mlpack" as the package name.

Julia

Get via Pkg

import Pkg;
Pkg.add("mlpack")

Ubuntu, Debian, Fedora, Red Hat

Install using a package manager

• mlpack command-line quickstart guide

Build from source

• Linux build tutorial

Windows

Download prebuilt binaries

• Windows 64 bit - MSI Installer
• Windows 64 bit - ZIP

Install using a package manager

• Installation through vcpkg

Build from source

• Windows build tutorial

macOS

Install using homebrew

brew install mlpack

Build from source

• Linux build tutorial